مدلسازی تغییرات کاربری اراضی فیروزآباد با استفاده از تصاویر ماهواره ای چند زمانه

Introduction

Today, studies and researches based on land use are carried out in order to meet the needs of the land surface with multiple uses such as agricultural land, gardens, residential areas and industrial areas, etc. Using remote sensing images as a new tool in surveying land cover and land use types, their changes over time can be checked and confirmed. In today's world and with the growth of technology, satellite observations are very important for understanding land use patterns in large areas and detecting changes over time.

In recent years, the use of satellite images in determining land use and investigating the expansion of cities has been of great interest. The use of remote sensing technology with the aim of examining changes over time is an inevitable necessity (Rosouli et al., 2015). Paying attention to the adverse effects of land use change in the management of urban areas by planners using Landsat satellite image information (Zairi, Amirani et al., 2017) is very useful. Therefore, due to the rapid growth of urbanization, not only a large amount of urban natural lands have changed their use, but the expansion of cities has caused socio-economic and physical use changes in the surrounding villages (Akbari 2012). Therefore, it can be claimed that the science of remote sensing is a comprehensive tool for managing natural resources (Qurbani et al., 2013) and remote sensing satellites are the best and most accurate data sources for detecting, quantifying and mapping patterns of land use changes (Andei Kawy et al. 2016).

Methodology

First, satellite images of three appropriate time periods were collected from different sources. The data were checked for geometric and radiometric errors and were received at the L1T level. To prevent errors, by displaying individual bands and different color combinations on the computer screen, the data in terms of radiometric errors were investigated and the atmospheric correction was done in the software using the ENVI method.

To check the geometrical situation, the georeferenced image was prepared from the USGS site, and the intersections and important points were checked in the ENVI software. Then, the prepared images were classified by image classification and band combination software such as ENVI 5 and ILWIS and based on ground control, the classification was done. Relevant corrections were applied according to the needs of the obtained images. Coding and combination of color images for different decades were carried out. The next step was to determine the number of desired classes for supervised classification and to determine land use, and then to ensure the accuracy of the classification, a scattered sample was collected from all the studied areas. The fuzzy logic method based on the theory of adaptability was used to classify the image and extract land use and land cover maps. Finally, using this model, a map of land use was prepared from the studied area, and to evaluate the changes, the area of each of the land use classes in each period was drawn in the form of a table.

To figure out the changes in the studied area, the maps obtained from the classification of images were studied with the orthogonal table method, and the prepared user maps were compared two by two and created as a matrix. The simulation was done in the relevant software such as IDRISI, GIS. By comparing the land use of the first two decades with the third decade and using the Markov chain method, simulating and predicting the trend of land use changes for the coming years was considered as the forecasting horizon in this research.

Result and discussion

The best band composition

Geometrical corrections were made in ENVI software. The obtained images had no significant error and atmospheric correction was also applied. The RGB composition of the image was prepared and the desired OIF index was calculated by ELWISI software to prepare the best band composition of the images. The obtained combination was pasted together in ENVI and the desired range was classified. In addition, the images received from TM and ETM were combined using spectral bands and panchromatic bands, which have high spatial resolution, and a better image was created.

Result Classification

The images were classified with four classes of residential areas, irrigated and rain-fed agricultural lands, and barren lands with two methods of supervised maximum likelihood and object-oriented classification. According to the size of the region and the geometric shape of agricultural lands and urban areas, more object-oriented images were considered.

Accuracy of Image Classification

To evaluate the accuracy of the classification, suitable locations for testing and different from the location of the specified training or user samples were necessary, and then the classified images were compared with the ground reality data in an error matrix.

Conclusion

The 15-year study period with the expansion of built-up lands, shows the most destruction in the agricultural lands of the outskirts of the city and barren lands in the border and suburbs of the city and agricultural lands. This trend can lead to environmental damages and further reduction of agricultural lands in the region. Previous research and experiences show that the use of a GIS system in the field of implementation and management of land use changes, especially agriculture, is inevitable. According to the result of the artificial neural network and Markov model and drought in recent years, it is necessary to note and manage the hydrology of the region. Also, due to the passage of two main roads, one from the southern industrial area (Southern Parian) and the other from the east, it is necessary to plan more precisely for the land uses and development of industrial and urban areas. The amount and type of cultivation in the plain lands of Firozabad need to be reconsidered, so that foreseeable problems in the region can be solved before they occur.

Undoubtedly, it is necessary to have a spatial database of the land use areas of the city under study, which comes from various sources, including satellite images, public information, and cadastral maps.